| Agricultural high-tech is a pioneering new technology group that takes agricultural application as the main body and has the conditions of intensive technology,intelligence and R&D funds.It can bring high economic growth rate and enhance international competitiveness.The intelligent evaluation of technology refers to the intellectualization of technology evaluation.The specific meaning is to apply machine learning and artificial intelligence technology to all aspects of technology evaluation,to promote the innovation of evaluation technology,methods and means,and to support the overall improvement of evaluation intelligence level with technology.Under the background that the country has comprehensively pushed forward the Supply-Side Reform,promoted economic transformation and put forward the development strategy of "Rrural Revitalization",China's independent innovation capability in the field of agricultural science and technology has been continuously improved,and a large number of agricultural high-tech have been continuously emerging,thus making a technological reserve for accelerating the modernization of agriculture and rural areas.How to objectively and scientifically evaluate the practical value of agricultural high-tech from the aspects of economic benefits,social benefits,environmental benefits,etc.to ensure and supervise the quality of technologies is an urgent problem to be solved to promote the transformation of scientific and technological achievements and the healthy and rapid development of agricultural and rural economy.In the process of technical evaluation,the evaluation report contains many contents and is highly professional.Writing the evaluation report is a complicated and tedious job.Using information technology and natural language processing technology to complete the online automatic generation of the evaluation report will greatly improve the efficiency of technical evaluation.This paper mainly discusses the theory and method of intelligent comprehensive evaluation of agricultural high-tech and adopts research methods such as literature research,index system,model analysis,experimental simulation and verification,and mainstream technology website research.In the exploration of the theory and practice of intelligent comprehensive evaluation of technology,the following aspects are mainly studied:(1)The definition and classification of agricultural high-tech are given.According to the characteristics of agricultural high-tech,the definition is given.In order to classify agricultural high-tech,this paper refers to the classification of agricultural science in the Chinese library classification and FAO agricultural thesaurus,and investigates the most widely used fields of agricultural high-tech in recent years in combination with the scope specified in "High-tech Fields Supported by the State".(2)A comprehensive evaluation index system of agricultural high-tech has been established.The selection principles and scope of various agricultural technology evaluation indexes at home and abroad are investigated.Starting from the characteristics of agricultural high-tech itself,the evaluation needs of stakeholders and the basic conditions for industrialization of agricultural high-tech,the comprehensive evaluation indexes of agricultural high-tech are screened,and the quantitative processing methods of the indexes are given.(3)The intelligent comprehensive evaluation model of agricultural high-tech is established.Select reasonable technical as samples,invite experts to score agricultural high-tech according to evaluation indexes as sample data,and use BP neural network to build an intelligent evaluation model.The evaluation effect of the model is compared with that of the model constructed by using AHP and TensorFlow deep learning framework,and get the conclusion that BP neural network has better evaluation effect is obtained.(4)The automatic generation model of technical evaluation report is constructed.The text in the evaluation report is divided into "technical text" and "evaluation text".The generation process adopts the extraction automatic summarization technology and the template generation method respectively,which improves the readability of the evaluation report and helps to summarize and analyze the evaluation results more objectively and efficiently.(5)A prototype system of intelligent comprehensive evaluation is designed and developed.The system consists of two parts: intelligent comprehensive evaluation and evaluation report generation.From the selection of technology,the recommendation of experts,the intelligent comprehensive evaluation of technology to the automatic generation of evaluation report,online operation is fully implemented,making the comprehensive evaluation of agricultural high-tech more automatic and intelligent.The main innovations of this paper are as follows:Firstly,the comprehensive evaluation index system of agricultural high-tech has been established,according to the characteristics of agricultural high-tech and its industrialization conditions and the evaluation needs of stakeholders,including 6 first-level indicators such as technological advantages and market prospects and 32 second-level indicators.This index system serves as the data collection basis for comprehensive evaluation of agricultural high-tech.Secondly,the comprehensive evaluation of agricultural high-tech is transformed into a multi-classification problem based on the machine learning theory,the BP Neural Network training model is used to fit the nonlinear relationship between evaluation index and technology type,and the Particle Swarm Optimization algorithm is used to optimize the initial weight of the model to improve the accuracy of model prediction.Thirdly,the automatic generation of evaluation report is realized,and the BERT-MMR combination model is constructed for automatic summarization of "technical text",which solves the problem of insufficient diversity of summarization generation in traditional TextRank algorithm. |
PDF Full Text Request